I’m Learning Not to Ask AI Just One Question

2 min read

A student writing at a laptop beside the steps Ask, Why, Options, Do nothing, Validate, Decide

While building Calyqen, I keep realizing how easy it is to ask AI a question and just accept the first answer.

That actually worries me a little.

AI can now help with things that would have taken much longer before.

It can look at a resume, compare it to a job, suggest changes, help with interview questions, or organize next steps.

That is really useful.

But I'm starting to think the most important part isn't only what I ask AI to do.

It is what I ask after it gives me an answer.

I'm trying to ask “why?” more

If AI tells me:

Change this part of your resume.

I don't want to just change it.

I want to ask:

Why?

What makes this better?

What information are you using?

What are you assuming?

Could changing it actually make something worse?

That extra question sometimes changes the whole answer.

And I want other options

Another thing I'm learning is not to ask AI for the answer.

I try to ask:

  • What are 3 other options?
  • What are the pros and cons of each?
  • What would someone disagree with here?
  • Is there a simpler way to do this?
  • What if I did nothing at all?

That last question is something I didn't really think about before.

Sometimes doing nothing is actually an option.

Maybe the resume is already fine.

Maybe a feature doesn't need to be built.

Maybe the problem isn't important enough to solve.

Maybe adding more technology actually makes something more confusing.

I think AI should help me see those possibilities too.

One word I'm starting to use a lot: “validate”

I don't want AI to just agree with my idea.

So I'm trying prompts like:

“Validate whether this is actually a problem before suggesting a solution.”

Or:

“Validate my assumptions. What might I be missing?”

Or:

“Give me reasons this idea might fail.”

Or even:

“Don't assume my idea is correct. Compare it with other options, including doing nothing.”

I actually like those answers more.

Sometimes they tell me something I don't want to hear.

But that is probably useful.

I still want the person making the decision

I don't think AI should decide if someone is qualified for a job.

I don't think it should decide what career someone should choose.

And I definitely don't think I should build a feature just because AI says it sounds like a good idea.

AI can help me think.

But I still want to decide.

Maybe that is the way I am starting to think about it:

Ask AI. Ask why. Ask for alternatives. Ask what happens if you do nothing. Validate the answer. Then decide for yourself.

I'm still figuring this out.

But the more I use AI, the more I think learning how to question the answer may be just as important as learning how to ask the question.

First published on LinkedIn.

See all